autoimmune NZB mice

Proc. Nati. Acad. Sci. USA

Vol. 76, No. 7, pp. 3464-3468, July 1979 Immunology

Abnormalities in clonable B lymphocytes and myeloid progenitors in autoimmune NZB mice (immunoglobulin genetics/IgD/hematopoiesis/colony-stimulating factors/prostaglandins)

PAUL W. KINCADE, GRACE LEE, GABRIEL FERNANDES, MALCOLM A. S. MOORE, NEIL WILLIAMS, AND ROBERT A. GOOD Sloan-Kettering Institute for Cancer Research, 145 Boston Post Road, Rye, New York 10580

Contributed by Robert A. Good, April 6, 1979

ABSTRACT Cloning procedures were used to study B lymphocytes and progenitors of granulocytes and macrophages in NZB mice. Numbers of B cells that were detected in sheep erythrocyte-containing semisolid cultures were only slightly elevated in NZB tissues, and these were normally sensitive to inhibition by anti-,u or anti-b antibodies or prostaglandin E. However, NZB mice rapidly developed large numbers of B cells that could be cloned in the presence of lipopolysaccharide, and these included unusual anti-,u resistant cells. Numbers of myeloid precursors in NZB bone marrow that were responsive to colony-stimulating activity in L-cell conditioned medium or endotoxin serum were at least normal, but at all ages granulocyte-macrophage precursors were poor responders in cultures stimulated by WEHI-3 cell conditioned medium. Almost no colonies were elicited in NZB cultures with a colony-stimulating activity moiety from WEHI-3 cells. Prostaglandin sensitivity of myeloid precursors from NZB and CBA mice was also different. Codominant genetic control of these abnormalities was suggested by their partial expression in F1 hybrid NZB X CBA and NZB X NZW mice. NZB mice expressed an unexpected IgD allotype allele.

ability of precursor cells of NZB mice to form myeloid colonies in response to a particular stimulus in vitro. Both of these were characteristic of young NZB mice of both sexes and were also partially expressed in F1 progeny of NZB mice mated with NZW or CBA strain mice.

Abnormalities have been found in or attributed to T cells, thymic epithelium, B cells, macrophages, and stem cells in autoimmune NZB mice, and they produce high levels of xenotropic virus (1-6). The relevance of each of these abnormalities to the development of the hemolytic anemia in these mice is not clear, but genetic studies tend to minimize the role of viruses and certain hemopoietic abnormalities in the etiology of disease (5, 6). A possible temporal sequence is suggested by findings that B cells of NZB mice are polyclonally activated by 1 week of age, and autoantibodies with preference for suppressor T cells are detectable early in adult life (7-10). A primary B-cell abnormality or defects that are expressed at the B-cell level could thus predispose to subsequent aberrant T-cell function. On the other hand, production of antierythrocyte autoantibody may occur in the absence of significant numbers of T cells, and potential for autoimmunity is transferrable with fetal liver or bone marrow cells from NZB mice (3, 11). Some gene products influencing manifestation of autoimmunity might therefore be expressed in pre-B cells, B cells, or nonlymphoid cells capable of activating B cells. Limiting dilution cloning techniques are available for enumerating and characterizing several hemopoietic precursors as well as lymphocytes, and we propose that these can be of value for identifying intrinsic defects in various cell types and for following changes in these with progression of disease. Thus far, we have found that particular culture conditions reveal an exceptional category of B lymphocytes in NZB lymphoid tissues, which, unlike B cells from many other strains, proliferate in the presence of anti-ji antibody. In addition, a severe abnormality was detected in the

MATERIALS AND METHODS Animals. Allotype congenic C57BL/6.Ige mice were donated by Noel Warner. NZB/Umc, CBA/H, (NZB X NZW)FI, (NZB X CBA)F1, and NZW/Umc were all produced at Sloan-Kettering from breeding stock obtained from the University of Minnesota. Cell Cultures. McCoy's modified 5a medium containing 15% fetal calf serum and enriched with respect to glutamine, asparagine, serine, pyruvate, and amino acids was used for all semisolid agar cultures as described (12-15). B lymphocytes were cloned in the presence of 50 ,gM 2-mercaptoethanol and either 25 jig of Salmonella typhosa endotoxin (lipopolysaccharide; LPS) per ml or 0.1% washed sheep erythrocytes (SRBC). The importance of using LPS or SRBC for achieving reproducible colony numbers and a linear dose-response curve is discussed in previous reports (12, 14). Cultures were examined with a dissecting microscope for the presence of colonies (aggregates of >20 cells) after 6 days of incubation at 370C in a fully humidified atmosphere of 7% CO2 in air. Granulocytic macrophage progenitors were similarly cloned in the absence of 2-mercaptoethanol in cultures containing an appropriate source of colony-stimulating activity (CSA). Granulocyte-Macrophage CSA. Stimuli used for cloning neutrophil and macrophage colonies were concentrates of conditioned medium from WEHI-3 myelomonocytic leukemia cells (WEHI-3-M) and L cells (LCCM). Both cell lines were propagated in 75-cm2 plastic flasks in McCoy's 5a medium containing 2.0% fetal calf serum and 50,uM 2-mercaptoethanol. Medium was harvested at 3- to 4-day intervals and concentrated 5-fold by Amicon (Lexington, MA) UM 10 ultrafiltration. Serum-free medium conditioned by WEHI-3 cells for 3-4 days was concentrated and equilibrated with 0.02 M Na phosphate buffer, pH 8.0, and then passed through DEAE-Sephadex as described (16). This material will be subsequently referred to as semipurified WEHI-3 CSA. A fourth source of CSA was serum from CBA or NZB mice taken 3 hr after an intravenous injection of 5 ,ig of LPS endotoxin. Prostaglandin E (PGE) was donated by John Pike (Upjohn), and this was stored at 1 mM in ethanol at -70°C. Anti-Immunoglobulin Antibodies. Anti-IgM antibodies were collected from goat anti-mouse myeloma M 104E (ji, X) serum on an absorbent of HPC-76 (ji, K) and rendered specific

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U. S. C. §1734 solely to indicate this fact.

Abbreviations: CSA, colony-stimulating activities; ES, serum from mice injected with endotoxin; LCCM, medium conditioned by L cells; LPS, lipopolysaccharide; PGE, prostaglandin E; SRBC, sheep erythrocytes; WEHI-3-M, medium conditioned by WEHI-3 cells.

3464

Immunology:

Proc. Natl. Acad. Sci. USA 76 (1979)

Kincade et al.

for it-chains by repeated passage over HOP-1 (y2a, A), UPC 10 (QY2a, K), and MPC 11 ('Y2b, K) immunoabsorbents. Goat anti-K antibodies were prepared by immunizing with F(ab')2 fragments of MOPC-460 (a, K), eluting the resulting antibodies from the HPC 76 column, and then crossabsorbing on the HOP 1 column. Polyvalent anti-Ig antibodies from rabbit anti-TEPC 183 (w, K) serum were eluted from a protein A-Sepharose (Pharmacia) absorbent. Ascites or serum from mice bearing the H6/31 anti-IgD allotype hybridoma (17) was purchased from Sera Labs (Crawley Down, Sussex, England). These reagents specifically lysed clonable B cells from C57BL/6J tissues at a 1:50 dilution in the presence of rabbit complement (1:12) and inhibited B-cell cloning when added directly to the cultures at a 1:200 final concentration. Under the same conditions, cloning of NZB or C57BL/6.Ige B cells was not significantly inhibited. Two other anti-IgD hybridomas produced at Stanford (18) were obtained from the Salk Institute Cell Distribution Center and maintained in RPMI 1640 medium. Culture supernates from the 11-6.3 hybrid cell line inhibited B-cell colony formation from C57BL/6 but not A/J, CBA/H, NZB, or C57BL/6.Ige mice at a 1:100 final dilution in the cultures. The pH of the 10-4.2 culture supernates was increased to 8.6 with concentrated Tris, and the supernates were passed through the protein ASepharose column. Approximately 3 ,tg of protein per ml of culture fluid was eluted from the column with glycine.HCl buffer, pH 2.8, and this was neutralized, concentrated, and dialyzed. This preparation contained only murine IgG2a by Ouchterlony analysis and was subsequently referred to as pure 10-4.2 anti-6. It was not cytotoxic for B cells with our rabbit complement but inhibited cloning of B cells from NZB, C57BL.Ige, and CBA/H mice when added directly to the cultures. B cells of both A/J and C57BL/6 strains are negative for this specificity. A broader discussion of the expression of IgD allotype alleles will be presented elsewhere (J. Goding, P. W. Kincade, and L. A. Herzenberg, unpublished results). Depletion of B Lymphocytes. B cells from spleen, lymph nodes, or bone marrow were depleted by adherence to anti-Ig coated petri dishes according to the method of Wysocki and Sato (19). Nonspecific cell loss was estimated by use of parallel uncoated dishes and by incubating allotype-negative cells in anti-6 coated dishes. Depletion of sIg+ cells always exceeded 97% when anti-,t or anti-IgM coated dishes were used. RESULTS Incidence of Clonable B Cells. When B cells are cultured in semisolid agar cultures in the presence of 2-mercaptoethanol and either LPS or SRBC, a linear relationship exists between the number of B cells present and the number of proliferating clones (14). Numbers of clonable B cells in NZB tissues were markedly elevated relative to the numbers and proportions of these found in CBA mice. This was apparent in LPS-potentiated cultures of mice 1-50 weeks old (Fig. 1). The incidences of colony-forming B cells in SRBC-potentiated cultures of CBA and NZB tissues were usually comparable, and the total numbers of colony-forming cells per spleen were consistently high in NZB mice only after development of splenomegaly (not shown). The numbers of colonies obtained with LPS alone or SRBC alone are not always equivalent with various normal B-cell populations, and approximately additive values are often obtained when the two potentiators are used together (12, 14). This was also true for NZB cells (data not shown). IgD Expression on NZB B Cells. A proportion of the colony-forming B cells in tissues of normal adult mice are extremely sensitive to anti-b antibodies (15). Others have found that the B cells of NZB mice have a low density of surface IgD relative to IgM (20), so we tested the effect of addition of anti-6 antibodies to cultures of NZB cells. An unexpected finding was

3465

lo" I

6ea

c a,

a,

* 0

'a

0 0

CD

Ca

0)

a

0o

e,

~~8

0o

5

10

15

20'25 30

35 40 45 50

Age, weeks FIG. 1. Clonable B lymphocytes in NZB (e) and CBA/H (0) spleens in cultures potentiated with LPS. Colony-forming units per

spleen were calculated by multiplying numbers of total nucleated cells per spleen times the average incidence of colonies in four replicate cultures of 2 X 104 cells.

that NZB B cells were completely resistant to H-6/31 hybridoma anti-6 antibody, whereas colony growth of A/J B cells was inhibited. These strains have been found to share alleles at all other immunoglobulin allotype loci (21). Furthermore, A/J B cells do not express the IgD specificity detected by the 10-4.2 hybridoma antibody, whereas both NZB and normal allotype congenic C57BL/6.Ige strain mice do. C57BL/6.Ige mice were derived from mating of NZB and C57BL/6 strains with subsequent backcrossing to C57BL, and BL/6.Ige mice share all known Ig alleles with NZB mice (22, 23). These results suggest that a recombination occurred such that alleles of the Ig-5 locus differ in A/J and NZB mice. In several experiments with cultures of NZB cells, both the degree of colony inhibition and the amount of purified 10-4.2 hybridoma anti-6 antibody required were similar to results obtained with other strains of mice (Fig. 2). The same result was obtained when heterologous rabbit anti-mouse IgD serum was used (data not shown). Sensitivity of Colony-Forming B Cells to Anti-,g Antibodies. We reported that inclusion of divalent anti-Ai or anti-K antibodies in LPS- or SRBC-potentiated semisolid cultures almost completely prevents colony formation (15). This has been true for cultures of B lymphocytes from CBA/H, CBA/J, CBA/cum, C57BL/6, A/J, B6D2FI, BALB/c, C3H, and SJL mice. A significant proportion of the LPS-potentiated colonyforming cells in NZB tissues are unaffected by 25 ,tg of anti-,u antibodies per ml. These unusual cells were detectable in spleen as early as 1 week after birth and reached maximal incidence at around 17 weeks of age (Fig. 3). At this time they comprised an average of 18%, 30%, and 41% of the LPS-potentiated colonies in culture of NZB bone marrow, spleen, or lymph nodes, respectively (Table 1). Anti-,u antibodies inhibited colony formation by greater than 99% in cultures of CBA tissues at all ages tested. Inhibition of NZB colony formation was essentially complete when polyvalent anti-IgM (,, K) was added to the cultures. Furthermore, an equivalent number of NZB and CBA spleen- or lymph node-clonable B cells adhered to petri dishes coated with anti-,u or anti-IgM antibodies (data not shown). Nonspecific cell loss with the dish depletion technique has always been less than 30% and is often less than 5%. This finding suggests that the anti-,I resistant cells in NZB tissues are in fact B cells and must express at least small amounts of surface IgM. In contrast to the results with LPS-potentiated cultures, both

Immunology: Kincade et al.

3466


-

0

0 on

0 0 0

).0 gg/ml FIG. 2. Sensitivity of NZB (0) and normal allotype congenic C57BL/6.Ige (0) B-lymphocyte colony formation to pure 10-4.2 hybridoma anti-6 antibodies. Each point is the mean (±SEM) number of colonies in four replicate cultures.

Table 1. Sensitivity of B lymphocyte colony formation to anti-p antibodies Mean no. of colonies ± SEM Medium alone Anti-,p NZB Bone marrow 801 i 147 145 + 51 Spleen 5417 + 601 1641 + 218 Lymph nodes 3103 + 192 1257 + 94 CBA Bone marrow 655 + 56 1± 1 Spleen 1059 + 87 7± 3 Lymphnodes 475 + 172 1+ 1 Sensitivity was measured in LPS-potentiated cultures with or without 25,ug anti-p antibodies per ml. Five 17-week-old NZB mice are compared to the same number of 9-week-old CBA mice.

Final concentration,

NZB and CBA B-lymphocyte colony formation in the presence of SRBC was completely inhibited by anti-A antibodies. Granulocyte-Macrophage Progenitors. Nonlymphoid colony formation in semisolid cultures is dependent on CSA obtainable from various sources. We employed WEHI-3-M for this purpose in the majority of our experiments. The incidence of WEHI-responsive progenitors was relatively constant at all ages in CBA bone marrow and averaged 162 + 8 colonies per 105 cultured cells in 23 determinations. NZB bone marrow was markedly deficient in these cells at all stages and averaged only 43 + 6 colonies per 105 cultured cells in 22 assays. The gross morphology of NZB colonies was abnormal in that the cells were unusually dispersed. This defect in in vitro myelopoiesis was even more apparent when a semipurified fraction of WEHI-3 CSA was employed and almost no NZB colonies were stimulated by it (Fig. 4). Cultures of equal part mixtures of NZB and CBA bone marrow gave intermediate numbers of colonies with this stimulator (data not shown). In striking contrast to these results with WEHI-3 CSA, NZB marrow cells gave at least normal numbers of colonies in cultures stimulated by LCCM or ES CSA, and the plateau of colony numbers normally seen with high CSA concentrations was not observed (Fig. 4). Prostaglandin Sensitivity of Colony Formation. Previous

.

studies from this laboratory have emphasized the regulatory potential of prostaglandins for lymphohemopoietic cells (24). B-lymphocyte cultures of NZB and CBA cells prepared with LPS or SRBC were similarly sensitive to 0.01-50,tM synthetic prostaglandin E (PGE) (data not shown). However, NZB colonies stimulated by WEHI-3 CSA were unusually sensitive to PGE, whereas their L-cell CSA-stimulated colonies were more resistant to PGE than were those of CBA controls (Fig. 5). Heritability of NZB Abnormalities. In most of our experiments we have compared NZB to normal CBA/H mice. It is important to establish whether the abnormal features of these autoimmune mice are common to the more closely related normal NZW mice and whether they are expressed in F1 progeny of NZB parents. One such comparison is illustrated in Fig. 6. NZW mice had more B cells that were anti-ti resistant in LPS-potentiated cultures than CBA mice but far fewer than NZB. Individual (NZB X NZW)F1 animals tended to have intermediate numbers of abnormal B cells, and the same was found for hybrids of NZB and CBA. The incidence of WEHI-3 CSA-responsive cells in NZW and CBA bone marrow were similar as were their sensitivities to 0.1 MAM PGE (Fig. 7). Again, F1 offspring tended to have characteristics that reflected the influence of both of the parental strains. Our experiments were not designed to discern small sex-related differences, but at least one of each type of analysis was done with males and females, and it seemed that the sex of the animals did not appreciably influence the results.

300f

104

0

U c 0

U

g 250 0

IA

i

4)

'a

EE 200 c

0 I

0 0

cD

VZ

3

* 150

0 .

; 0

d

103:7

S

C

a

._

0

100IF

/0

d

50 /

o

d1

U

102-

1:64 1:16 1:4 5

10

15 20 Age, weeks

25 30 35 40 45

FIG. 3. Incidence in spleen of NZB mice of B cells that form colonies in LPS-potentiated cultures containing 25 pAg of purified anti-p antibodies per ml.

un-1:64 1:16 1:4 dil. CSA dilution

un-

dii.

FIG. 4. Differential stimulation by various granulocyte-macrophage CSA sources. Dilutions of semipurified WEHI-3-M (v), LCCM (-), or endotoxin-stimulated serum (ES) (0) CSA prepared as described in Materials and Methods were added to cultures of NZB (Left) or CBA (Right) bone marrow cells.

Immunology:

Kincade et al.


'aG

u

10-8

200

E 200

I~~~~~~I

C

0

.0 -o

*a 150

*

150

S~~~~~~~~

* o 100

.

-A

0

*'D0

10-6

0 C.

50

~

I

_1

0

: -T ~~~~~~~~~~~

100

-

50

i4+

Sensitivity of granulocyte-macrophage colony formation

5.

@)

of CBA/H

and NZB (I) cells to PGE. In this experiment CBA/H

and NZB bone marrow formed 150

5 and 55

i

3 colonies per

105

cultured cells, respectively, in unfractionated WEHI-3-M stimulated cultures (Left) containing no prostaglandin. In LCCM-stimulated cultures (Right) the uninhibited control values were and 185

250

6,

PGE, M

FIG.

CSA + 0.1 AM PGE

250

0L-

0~

-06

CSA alone

2

3467

i

88+i5 for CBA/H

3 for NZB cells.

+

--

FIG. 7. Incidence of granulocyte-macrophage progenitors and their sensitivity to prostaglandin in parental and hybrid progeny mice. Assays were done with the same individuals shown in Fig. 6 and unfractionated WEHI-3-M as the source of CSA.

DISCUSSION The two salient findings of this study were: (i) an exceptional type of B cell, which proliferates in LPS-potentiated semisolid

cultures and resists anti-g antibodies, exists in NZB lymphoid tissues, and (ii) defective responsiveness of hemopoietic precursor cell populations to certain types of colony-stimulating activity was demonstrated. These abnormalities were partially expressed in F1 hybrids with two different strains and thus may be under codominant genetic control. The B cells that proliferate to form colonies in agar cultures have been extensively characterized. Clonable cells are almost as heterogeneous as all other B cells in terms of physical properties and other respects, and the only population of slg+ cells that have been found to totally lack cloning potential are those

C

a,a

106

-

.

()

c

.r,

.sE

10

0

0

c

0

104

io4

4-1

4-§

q-

K-

C,

11,v~o

FIG. 6. Heritability of anti-j resistant B colony-forming cells in F1 hybrid offspring of NZB mice. Individual values together with means for each group are given for 17-week-old NZB, 18-week-old NZBxNZW, 18-week-old CBAxNZB, 10-week-old NZW, and 9week-old CBA/H mice.

present in partially immunodeficient CBA/N mice (13). The minimal culture requirements for colony formation are 2mercaptoethanol, fetal calf serum, and appropriate mitogens such as those native to laboratory grade agar (12). Modification of these basic conditions, such as the use of macrophage feeder layers, addition of LPS, or addition of SRBC are done to optimize colony size and numbers and, most importantly, to achieve consistent, linear dose-response curves (14). Indirect observations suggest that the populations of B cells that respond under these different conditions may not be completely overlapping. More colonies develop in cultures in which more than one potentiator is used, and occasionally this number equals the sum obtained with each one used alone (12, 14). Adult bone marrow B cells have a higher cloning efficiency in LPS- than in SRBC-potentiated cultures (15, 25). NZB mice develop very large numbers of cells that clone in LPS-containing cultures, whereas the incidence of clonable cells in SRBC-containing cultures was usually similar to that of normal CBA/H mice. Others have found that B cells in very young NZB mice are polyclonally activated to differentiate and secrete immunoglobulin, and the autoimmune potential has been transferred to irradiated normal recipients with cells other than T lymphocytes (3, 7, 8). These findings suggest that abnormal B-cell regulatory mechanisms could be important in the pathogenesis of this disease. We found that clonable B cells from NZB mice were normally susceptible to inhibition by PGE and anti-6 antibodies. In the latter case, nanogram quantities of purified antibodies achieved a plateau of inhibition, and in other studies we determined that the anti-6 resistant B cells include some cells that are surface 6- and others that are 6+ (unpublished results). Our results thus do not contradict other reports that NZB B cells have a low density of cell surface IgD relative to IgM (20), but indicate that the 6 receptors on their clonable cells function normally in regulation. As predicted from all of our previous studies, the B cells of NZB mice do not clone in anti-cs containing, SRBC-potentiated cultures. However, a substantial proportion of the LPS-stimulated clones resist anti-1 suppression. These cells were completely sensitive to inhibition by polyvalent anti-Ig and thus are B cells, and, because they were depleted by incubation in anti-g or anti-IgM coated petri dishes, they may actually express sIgM. The maximum incidence of these unusual B cells was found in adult NZB lymph nodes, and this would indicate that they belong to relatively mature and perhaps memory cell populations.

3468

Immunology: Kincade et al.

Our fortuitous observation that NZB and A/J mice differ in IgD allotypes is of interest with respect to immunoglobulin genetics. These mice share other Ig allotypes, and recombination of Ig allotype genes is an extremely rare event in the mouse (21). The Ig-5 (IgD) locus could be distal to the other Ig allotype genes but must nonetheless be linked closely to them because NZB and C57BL/6.Ige mice share all allotype alleles (22, 23). Where lymphocytes require mitogens for clonal proliferation, precursors of eosinophilic, megakaryocytic, erythroid, granulocytic, and monocyte-macrophage cells require factors collectively referred to as CSA for colony formation in vitro (26, 27). A wide variety of normal and neoplastic cells elaborate CSA, and the diversity of these substances, their nature, and the relationship between these different molecules is far from understood. Myeloid progenitors in NZB marrow were almost totally refractory to a CSA moiety in semipurified WEHI-3-M over a wide concentration range. In contrast, at least normal numbers of colonies were obtained with LCCM or ES CSA, and it is interesting that numbers of NZB colonies did not plateau with high concentrations of these stimulators. This perhaps explains the elevated numbers of colonies previously observed with NZB cultures containing ES (28). In addition to these abnormalities, the sensitivity of NZB precursors to PGE was different from those in normal CBA marrow. The different CSA are known to differ in terms of the morphology of the colonies they elicit, but this does not correspond to the NZB abnormality. Most preparations of semipurified WEHI-3-M preferentially stimulate neutrophil colonies and LCCM elicits almost pure macrophage colonies (16, 29). However, NZB precursors made at least normal numbers of granulocyte colonies in ES and failed to respond to batches of WEHI-3-M that were significantly contaminated with monocyte-macrophage stimulators. Further study is required to determine whether one or more subsets of myeloid precursors is absent in NZB marrow or whether these cells are defective in their recognition of certain regulators. We have also found (unpublished data) that (i) NZB marrow cells survive and function poorly under continuous marrow culture conditions that permit proliferation and differentiation of cells of other mouse strains for many weeks, (ii) abnormalities in in vitro myelopoiesis are apparent with early fetal liver cells and can be transferred to lethally irradiated BDF1 recipients, (iii) NZB precursors are abnormal in responsiveness to certain other regulatory substances, and (iv) these phenomena are probably due to intrinsic cellular abnormalities rather than active suppressor mechanisms. None of the unusual features that we describe here for NZB mice are necessarily involved in their autoimmune disease. Further study is also required to determine if and how polyclonal B cell activation events may be related to nonlymphoid abnormalities. In the meantime NZB mice may prove useful for dissecting hemopoietic precursor cell populations and the macromolecules that regulate their function. We are grateful to Drs. J. Goding and N. Warner for providing encouragement, advice, reagents, and mice. Dr. P. Ralph maintained the hybridoma cell lines and made helpful comments on the manuscript. We thank Mr. R. R. Eger for providing unique preparations of semipurified GSA. This work was supported by Grants AI-12741, AI-11843, CA-17404, CA-08748, AG-00541, and Research Career Development Award AI-00265 from the U.S. Public Health Service, and by the National Foundation-March of Dimes.

Proc. Nati. Acad. Sci. USA 76 (1979)

Krakauer, R. S., Waldman, T. A. & Strober, W. (1976). J. Exp. Med. 144, 662-673. 2. Cantor, H., McVay-Boudreau, L., Hugenberger, J., Naidorf, K., Shen, F. W. & Gershon, R. K. (1978) J. Exp. Med. 147, 11161.

1125. 3. DeHeer, D. & Edgington, T. S. (1978) J. Immunol. 118, 1858-1863. 4. McCombs, C., Hom, J., Talal, N. & Mischell, R. I. (1975) J. Im-

munol. 115, 1695-1699.

5. Warner, N. L. & Moore, M. A. S. (1971) J. Exp. Med. 134,

313-334.

6. Datta, S. K., Manny, N., Andrzejewski, C., Andre-Schwartz, J. & Schwartz, R. S. (1978) J. Exp. Med. 147,854-871. 7. Moutsopoulos, H. M., Boehm-Truitt, M., Kassan, S. S. & Chused, T. M. (1977) J. Immunol. 119, 1639-1644. 8. Izui, S., McConahey, P. J. & Dixon, F. J. (1978) J. Immunol. 121, 2213-2219. 9. Klassen, L. W., Lynell, W., Krakauer, R. S. & Steinberg, A. D.

10.

(1977) J. Immunol. 119,830-837. Shirai, T., Hayakawa, K., Okumura, K. & Tada, T. (1978) J.

Immunol. 120, 1924-1929. 11. Morton, J. I. & Siegel, B. V. (1974) Proc. Nati. Acad. Sd. USA 71, 2162-2165. 12. Kincade, P. W., Ralph, P. & Moore, M. A. S. (1976) J. Exp. Med. 143, 1265-1270. 13. Kincade, P. W. (1977) J. Exp. Med. 145,249-263. 14. Kurland, J. I., Kincade, P. W. & Moore, M. A. S. (1977) J. Exp. Med. 146, 1420-1435.

15. Kincade, P. W., Paige, C. J., Parkhouse, R. M. E. & Lee, G. (1978)

J. Immunol. 120, 1289-1296. Williams, N., Eger, R. R., Moore, M. A. S. & Mendelsohn, N. (1978) Differentiation 11, 59-3. 17. Pearson, T., Galfre, G., Ziegler, A. & Milstein, C. (1977) Eur. J. 16.

Immunol. 7,684-690.

18. Oi, V. T., Jones, P. P., Goding, J. W., Herzenberg, L. A. & Herzenberg, L. A. (1978) in Lymphocyte Hybridoma8, eds.

Melchers, F., Warner, N. L. & Potter, M. (Springer, Basel, Switzerland). 19. Wysocki, L. J. & Sato, V. L. (1978) Proc. Natl. Acad. Sci. USA 75,2844-2848. 20. Cohen, P., Ziff, M. & Vitetta, E. S. (1978) J. Immunol. 121, 973-977. 21. Herzenberg, L. A. & Herzenberg, L. A. (1978) in Handbook of

Experimental Immunology, ed. Weir, D. M. (Blackwell, Oxford), 3rd Ed., pp. 12.1-12.23. 22. Coding, J. W., Warr, G. W. & Warner, N. L. (1976) Proc. Nati. Acad. Sci. USA 73,1305-1309. 23. Goding, J. W., Scott, D. W. & Layton, J. E. (1977) Immunol. Rev. 37, 152-186. 24. Kurland, J. & Moore, M. A. S. (1977) Exp. Hematol. 5, 357373. 25. Lala, P. K., Johnson, G. R', Battye, F. L. & Nossal, G. J. V. (1979)

J. Immunol. 122,334-341.

26. Metcalf, D. (1977) In Vitro Cloning of Normal and Leukemic Cells (Springer, New York). 27. Burgess, A. W., Metcalf, D. & Russell, S. (1978) in Differentiation of Normal and Neoplastic Hematopoietic Cells, Cold Spring Harbor Conferences on Cell Proliferation, eds. Clarkson, B., Marks, P. A. & Till, J. E..(Cold Spring Harbor Laboratory, Cold Spring Harbor, NY), Vol. 5, Book A, pp. 339-357. 28. Metcalf, D. & Stevens, S. (1972) Cell Tissue Kinet. 5, 433446. 29. Stanley, E. R. & Guilbert, L. J. (1979) in Mononuclear Phagocytes-Functional Aspects, ed. van Furth, R. (Nijhoff, The Hague, Netherlands), in press.